Blue Clouds and Green Clouds: Virtual Biopsy via EUS Elastography?

The quality of “elasticity” is part of everyday life and is integrated into one of the senses - that of touch. We evaluate the morphology of a vegetable such as a tomato by rotating it between our fingers and pressing it to feel whether the texture is harder or smoother (and therefore overripe), to determine whether it is suitable for eating. In medicine, we use the fact that diseases change the elastic properties of organs. As early as the Ebers papyrus, dating from 1500 BC in Egypt, malignant tumors were described as “hard swellings.” The tactile sense of the physician’s hand was therefore used even in ancient medicine, whilst “percussion” and “auscultation” are relatively recent additions to the physician’s diagnostic armamentarium, dating from the end of the 18th century.

Auscultation developed into ultrasonography, and the elastic properties of the organs have been used in medicine since 1955, when elastography was developed to help detect thromboses [1]. Quantitative imaging of elastic modulus distributions in soft tissues was first reported as a tool complementary to ultrasound in 1991 [2]. The method was based on external tissue compression, with subsequent computation of the strain profile along the transducer axis. The strain profile was converted into an elastic modulus profile by measuring the stresses applied by the compressing device and applying certain corrections for the nonuniform stress field [2]. In recent years, various aspects of elastography have been further developed. It is now used to differentiate the mechanical properties of tissue and tissue pathology throughout the entire human body - for example, in the breast, prostate, thyroid, and brain [3] [4] [5] [6]. It has also resulted in new, commercially available devices such as the FibroScan (EchoSens, Paris, France) for assessing the stiffness of tissue in the liver [7]. In addition to its use in ultrasonography, elastography is also used to assess stiffness properties in magnetic resonance imaging and optical coherence tomography [8] [9].

Elastographic assessment is now also available in endoscopic ultrasonography (EUS), and Giovannini and colleagues are to be congratulated on conducting the first human trial of the method [10]. This is always a challenge, since the methodology and standards first have to be set with new tools and devices when their performance is not yet known. Fifteen years ago, an ill-defined hypoechoic appearance in a lesion on EUS led to the assumption that it represented malignancy, and lessons about the reliability of the finding first had to be learnt.

The first standard that Giovannini et al. set was the definition of a color code for the different qualities of hardness (elasticities) found for pancreatic lesions and lymph nodes. The second was to group the different elastographic findings (now color-coded) into various hardness patterns that could then be compared with the final diagnosis of the lesions. For breast cancer, different echogenic patterns have been defined: very dark and well-defined; moderately dark, with darker foci inside; or a very dark central core with a less dark peripheral component [3]. In this EUS-guided study, however, the color-coded elastic properties found were defined as: blue - malignant; green - fibrosis; yellow - normal tissue; red - fat. These scores directly relate the elasticity to the histology, and this would indeed represent a “virtual biopsy.”

However, this approach appears to be methodologically difficult, since a specific level of hardness, in this case indicated by blue, is set as equivalent to malignancy. This would lead us to believe that a malignant tumor is necessarily a hard entity that could easily be differentiated from green/fibrosis, for instance, while in fact we all experience in everyday practice the fact that hard lesions - for example, focal chronic pancreatitis - may not always be malignant, or that soft ones such as mucinous adenocarcinomas may not always be benign. If one looks closely at the illustrations provided, blue and green appear to merge into each other and no information is provided about how much green is allowed in a supposedly blue area for it still to be considered “malignant.” The rationale for why blue represents malignancy, etc., appears to be difficult to understand and is not further explained.

However, the notion of a virtual biopsy is attractive, and the authors were clearly aware of the shortcomings of the original simple, histology-related scoring system. They therefore developed a different and more easily understandable descriptive scoring system, shown in Figure 5 (page 345). Based on a review of experience subsequent to the study, this was added to the current results, and data validating the scoring system are not provided. The paper is thus essentially a preliminary pilot study attempting to develop criteria and standards, rather than testing established ones. However, it would be a self-fulfilling prophecy if the same group of patients were used both to develop criteria and standards and to test them.

Even the new scores are somewhat vague, and it would improve the method if consideration could be given to developing more objective means of obtaining such scores. This appears eminently possible - for instance, using a large number of patients with the same confirmed disease, or some form of computerized evaluation such as that used in densitometry, or both.

It is not clear whether the low specificity of 67 % for pancreatic lesions and 50 % for lymph nodes was due to problems with defining an objective color code or due to the nature of the lesions, their size, and the degree of difference between pathological tissue and the surrounding normal tissue. One of the latest reports on elastography in breast cancer reported wide variation in specificity - the overall specificity was 89 % (with a sensitivity of 79 %), with a specificity of 100 % for tumors < 2 cm in size dropping to 62 % in tumors > 2 cm [11].

Giovannini et al. have explored the potential of add-on elastography for differentiating between benign and malignant pancreatic lesions and lymph nodes, and they hold out the attractive prospect that it may be possible to develop a method of virtual biopsy. The color images do not yet provide sufficient clarity to be able to replace the information obtained from tissue diagnosis. However, this preliminary pilot study suggests that EUS elastography may be capable of developing into a useful diagnostic tool once it has been evaluated in greater depth in future trials including larger numbers of patients.

At present, the green clouds merge into blue clouds and blue clouds merge into green clouds (Figutre 3, 4, 6, 7 [10] and Figure [1] in this editorial) - but if and when the method can be improved, it may be able to provide guidance on where to biopsy, in the same way as it does in radiography, urology, and gynecology.

Competing interests: None

Figure 1 Vincent van Gogh (1853 - 90): The Starry Night (1889). Oil on canvas, 72 × 92 cm (Museum of Modern Art, New York).

References

• 1 Bay M. Thrombo-elastography: a new method for the study of thromboembolic disease; (in French).  Presse Méd. 1955;  63 241-242
• 2 Ophir J, Cespedes I, Ponnekanti H. et al . Elastography: a quantitative method for imaging the elasticity of biological tissues.  Ultrason Imaging. 1991;  13 111-134
• 3 Garra B S, Cespedes E I, Ophir J. et al . Elastography of breast lesions: initial clinical results.  Radiology. 1997;  202 79-86
• 4 Lorenz A, Ermert H, Sommerfeld H J. et al . Ultrasound elastography of the prostate: a new technique for tumor detection.  Ultraschall Med. 2000;  21 8-15
• 5 Lyshchik A, Higashi T, Asato R. et al . Thyroid gland tumor diagnosis at US elastography.  Radiology. 2005;  237 202-211
• 6 Scholz M, Noack V, Pechlivanis I. et al . Vibrography during tumor neurosurgery.  J Ultrasound Med. 2005;  24 985-992
• 7 Castera L, Vergniol J, Foucher J. et al . Prospective comparison of transient elastography, Fibrotest, APRI, and liver biopsy for the assessment of fibrosis in chronic hepatitis C.  Gastroenterology. 2005;  128 343-350
• 8 Muthupillai R, Lomas D J, Rossman P J. et al . Magnetic resonance elastography by direct visualization of propagating acoustic strain waves.  Science. 1995;  269 1854-1857
• 9 Khalil A S, Chan R C, Chau A H. et al . Tissue elasticity estimation with optical coherence elastography: toward mechanical characterization of in vivo soft tissue.  Ann Biomed Eng. 2005;  33 1631-1639
• 10 Giovannini M, Hookey L C, Bories E. et al . Endoscopic ultrasound elastography: the first step towards virtual biopsy? Preliminary results in 49 patients.  Endoscopy. 2006;  38 342-346
• 11 Guisseppetti G M, Martegani A, Di Cioccio B, Baldassarre S. Elastosonography in the diagnosis of the nodular breast lesions: preliminary report.  Radiol Med. 2005;  110 69-76